Incorporating on-line process data into a diagnostic knowledge-based system

Due to the character of the original source materials and the nature of batch digitization, quality control issues may be present in this document. Please report any quality issues you encounter to digital@library.tamu.edu, referencing the URI of the item.Includes bibliographical references.The objective of this research was to develop and implement an on-line gypsum wallboard manufacturing expert system incorporating on-line process monitoring information and statistical process control. In support of this objective, several sub-objectives were accomplished. First, an expert system addressing the major problems associated with gypsum wallboard manufacturing was developed. Second, manufacturing problems that utilized quantifiable process data in their problem solving procedures were identified for incorporation into the system. Third, the statistical process control techniques of control charting to filter process information were applied to selected problems in the manufacturing process. Finally, the system was implemented and verified. Several stagesand accompanying methods marked the development of this system. Identifying crucial parameters for expert system development, acquiring knowledge for those identified parameters, coding the knowledge base, adapting statistical process control (SPC) techniques to filter process data, and lastly, implementing the system in the study facility were the elements of the full expert system. Nine major categories of problems associated with gypsum wallboard manufacturing were identified and addressed by the expert system. Specifically, they were various types of problems associated with voids, the foam, bond, wallboard additives, the board line, the knife and inverter, the takeoff of the process end, and the warehouse. Also, the process data associated with each problem was identified, filtered using SPC methods, and incorporated into the problem solving process. The SPC techniques employed consisted of two different types of control charting and a data correlation technique. Individual charts and range charts were used as control charts. Autocorrelation analysis was used to determine data correlation. The expert system developed in this project has proven beneficial in documenting and preserving knowledge relating to gypsum wallboard manufacturing. The access to filtered process data supplied by SPC techniques built into the expert system provided the opportunity for a more objective and accurate assessment of the process related problems